32. Data-driven farming: how to read growth data (and make better decisions)

In indoor systems and modern vertical farming, growing without data is impossible.
Sensors, software and artificial intelligence generate an enormous amount of information, but the real competitive advantage is not collecting data: it is knowing how to interpret it correctly.

Data-driven farming is not an abstract concept: it is the ability to transform biological and environmental signals into operational decisions that can be measured, replicated, and improved over time.

In this article we look at what data really matters, how to read it, and how to use it to increase yield, quality, and production stability.

What data-driven farming really is

Data-driven farming is an approach in which:

• every growth parameter is measured

• every variation is tracked over time

• every decision is based on evidence, not hunches

In practice, the plant becomes a continuous source of data, not just a final output (the crop).

This approach is critical because:

• it reduces systemic errors

• makes production scalable

• enables continuous optimization

• enables artificial intelligence

The three levels of data in indoor cultivation

To read growth data correctly, it is essential to distinguish them by level.

1. Environmental data

These are the most immediate and most widely used:

• air temperature

• relative humidity

• CO₂

• air flows

• water temperature

These data describe the context in which the plant is growing, but they still do not tell how the plant is responding.

Common mistake: optimizing only the environment without observing the biological response.

2. Nutritional and water data

Here we begin to get into the physiology of the plant:

• EC

• pH

• water consumption

• nutrient uptake

• changes over time

An isolated datum is of little value.
Variation in the datum over time is what signals stress, excess or deficiency.

Example:

• Stable EC + slowing growth = non-nutritional problem

• Rapidly declining EC = plant in high metabolic activity

3. Real (biometric) growth data.

This is the most advanced and most underestimated level:

• growth velocity

• leaf development

• color and texture

• visual patterns

• uniformity among plants

This is where computer vision and AI come into play, enabling it to read signals invisible to the human eye before the problem becomes apparent.

It is on this level that data-driven farming becomes truly predictive.

The problem of "unarmed" data

Many plants collect data but do not generate value.

Why?

• unrelated data

• absence of history

• lack of reference models

• no feedback loop

The result is a dashboard full of numbers, but no automatic or suggested decisions.

Data become useful only when they are:

• contextualized

• comparable

• linked to the plant's response

From raw data to the Growth Plan

In evolved data-driven farming, the goal is not to "monitor," but to build and improve Growth Plan.

An effective Growth Plan:

• defines environmental and nutritional targets

• observes the plant's actual response

• automatically corrects parameters

• improves cycle after cycle

Here the data becomes operational, not just informational.

The role of artificial intelligence

AI does not replace the agronomist, but it does what the human cannot do:

• analyze millions of datapoints

• find correlations that are not obvious

• anticipate stresses and yield declines

• adapt parameters in real time

In the Tomato+ model, each greenhouse is a node that:

• collects environmental, nutritional and visual data

• sends it to the cloud

• helps train increasingly accurate growth patterns

The value is not in the individual plant, but in the network of distributed data.

Why reading data means reducing costs

A well-designed data-driven system enables:

• reduce energy waste

• avoid over-lighting

• optimize cycles

• prevent crop failures

• standardize quality

This is especially relevant today, where energy and operational stability are the real bottleneck in vertical farming.

The real paradigm shift

Data-driven farming changes one fundamental thing:

👉 you no longer grow a plant, you grow a growth model.

The plant becomes the physical validation of a digital system that learns, corrects and improves.

And this is the step that transforms a greenhouse from a "farming machine" to a scalable technology platform.

Conclusion

Those who do not read data, react to problems.
Those who interpret them correctly, prevent them.

In the vertical farming of the future, crop quality will depend less and less on the human hand and more and more on the ability to read, correlate and use growth data.

This is where the real competition lies.

Thank you for reading this article. Keep following us for new content on hydroponics, vertical farming, and smart agriculture.
Tomato+ Team